Missouri University of Science and Technology

Scholars' Mine
International Conference on Case Histories in
Geotechnical Engineering

(1993) - Third International Conference on Case

Histories in Geotechnical Engineering

Jun 1st, 12:00 AM

Building collapse due to shallow trench excavation

S.J. Vitton
R.C. Brown

Follow this and additional works at: http://scholarsmine.mst.edu/icchge

Recommended Citation
S.J. Vitton and R.C. Brown, "Building collapse due to shallow trench excavation" ( June 1, 1993). International Conference on Case
Histories in Geotechnical Engineering. Paper 1.
http://scholarsmine.mst.edu/icchge/3icchge/3icchge-session08/1

This Article - Conference proceedings is brought to you for free and open access by the Geosciences and Geological and Petroleum Engineering at
Scholars' Mine. It has been accepted for inclusion in International Conference on Case Histories in Geotechnical Engineering by an authorized
administrator of Scholars' Mine. For more information, please contact weaverjr@mst.edu.

!!

Proceedings: Third International Conference on Case Histories in Geotechnical Engineering, St. Louis, Missouri,
June 14, 1993, Paper No. 8.10

"'.;..:!'"

Building Collapse Due to Shallow Trench Excavation

S. J. Vitton

R. C. Brown

Assistant Professor of Civil Engineering, The University of

Alabama, Tuscaloosa, Alabama

Professor of Civil Engineering, The University of Alabama,

Tuscaloosa, Aalabama

SYNOPSIS: A building collapse is investigated in which a shallow trench, excavated along the length
of the building, resulted in the building's collapse. An investigation indicated that the trench had
been excavated within 1 ft (0.3 m) of the foundation and that the foundation had been constructed on
approximately 7 ft (2.1 m) of soft silty clay with the bottom portion of this soil highly saturated.
The bearing capacity of the foundation, prior to excavation, was estimated to have a safety factor as
low as 1.5.
As a result of the proximity of the trench to the foundation and the soft soils, the
trench collapsed. soon after excavation. It is believed that the lower portion of the soil flowed into
the tre~ch c~us~ng a loss . of be~ring capacity.
An understanding of the geologic setting of the
foundat~on so~ls, coupled w~th so~ls information from the USDA's Soil Conservation Service
indicated
the hazards of excavating in this soil.
'

INTRODUCTION

SITE CONDITIONS

on July 16, 1989 a shallow trench was excavated

along the length of a single-story building
located in the city of Tuscaloosa, Alabama. The
trench was being excavated for the installation of
a sewer line to provide service to a rear portion
of the building. The trench was approximately 3
ft ( 0. 9 m) wide, starting from the southwest
corner of the building and extending past the
northwest corner of the building as shown in
Figure 1.
It was estimated that the trench was
excavated to a depth of between 5 and 6 ft (1.5
to 1.8 m).
A
majority of the excavation was
completed in a four hour period by a single
backhoe operator.
The collapse of the building
occurred approximately one hour after the operator
completed the excavation and was on a break.
Fortunately,
no injuries resulted when the
building collapsed into the trench. A majority of
the west wall of the building rotated into the
trench resulting in the collapse of the building's
roof.
Ensuing litigation resulted in the contractor in
charge of the excavation being responsible for
damages.
Although
the
cause
and
effect
relationship of the collapse appeared obvious, the
extent and nature of the damage indicated that the
foundation soils were in a relatively weak state
prior to excavation. Consequently, a preliminary
study was conducted to determine the factors that
lead to the collapse of the building as well as
ways of identifying such sites that may be
susceptible to collapse from shallow trench
construction, thus helping to prevent similar
incidence.

The City of Tuscaloosa is located in the westcentral part of Alabama and has a population of
approximately 75,000. The climate in Tuscaloosa
is influenced by the Gulf of Mexico resulting in
relatively warm humid summers accompanied by
relatively mild winters. The average summer time
temperature is 80F 1 and the average winter time
temperature is 46 F with an average annual
precipitation of 52 in. The major physiographical
feature of the area is the Black Warrior River
Basin and the Black Warrior River.
The Black
Warrior River, which flows through Tuscaloosa,
provides a major transportation link with the Gulf
of Mexico and has two river lock systems within
the corporate limits of Tuscaloosa. The city is
located on the south bank of the Black Warrior
River on a terrace approximately 50 to 75 ft (15
to 23 m) above the river s maintained level of 123
ft MSL (37.5 m). Beyond the banks of the Black
Warrior River, Tuscaloosa tends from relatively
flat terrain in the West to relatively steep hills
to the East.
The collapsed building is located within a mixed
residential/commercial section of Tuscaloosa in a
southwest section of the city at the intersection
of two streets as is shown in Figure 1.
The
building was used by an electrical contractor as
both an office and warehouse at the time of the
collapse.
The area surrounding the building is
characterized as a topographically low area but
within a relatively flat section of Tuscaloosa.
The building was of cinder block construction
with conventional strip footings, which were
placed at grade level. The exterior walls of the
building were load bearing walls as well as a
center wall constructed along the length of the
building. The east part of the building, which is

Third International Conference on Case Histories in Geotechnical Engineering

Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu

1177

still in use as a storage building, has a 10 ft

(3.0 m) high flat roof, while the west one half of
the building was used as a warehouse with a 16 ft
( 4. 9 m) roof. The remaining concrete floor on the
west side of the building is now used as an
outdoor storage area. The footings were measured
at 16 in. (406 mm) in width and 18 in. (0.5 m) in
depth, while the combined bearing loads on the
footings were estimated to be 325 psf (16 kPa).
It is believed that the building was built in the
late 1940s or early 1950s.
NORTH

Parking Area

Figure 1

Plan view of the collapse building site.

GEOLOGIC CONDITIONS
Geologically, Tuscaloosa is located on the "fall
line" that separates the
Southern Appalachian
Mountains (Appalachian Plateau Province) and the
northern Gulf of Mexico coastal plains (Coastal
Plain Province).
The fall line is so named
because it marks the area where river navigation
from the Gulf of Mexico typically ends due to an
increase in gradient of the rivers in Appalachian
Mountain areas.
In this region the Appalachian
Plateau Province consists of relatively horizontal
Paleozoic sedimentary rocks, while the Coastal
Plain Province consists primarily of sedimentary
Cretaceous materials overlain with younger
Tertiary and Quaternary soils and rocks (Adams, et
al., 1926).
The surface geology of the Tuscaloosa area
consists
of
a
very
young
unconsolidated
sedimentary layer believed to be of Pleistocene
age and described locally as "terrace deposits".
These terrace deposits are alluvial deposits from
the nearby Black Warrior River and in the
Tuscaloosa area
lie unconformally on the
Pottsville Formation of Lower Pennsylvanian age
(Wielchowsky, 1975) . These deposits are believed
to have be~n deposited during the melting of the
great continental ice sheets during the final
phases of the last ice sheet (Wisconsin) some 10
to 30 thousand years ago. The terrace deposits,
which are 40 to 100 ft (12 to 30 m) thick in the
Tuscaloosa area, are marked by coarse gravel at
the bottom and become less gravelly and more sandy
vertically until the upper layers are largely fine
sand. The upper half is sandy in the lower part,
yielding to red clayey soils nea~ the top . T~e
red clayey soils are characterist1c of later1t1c
weathering (Adams, et al., 1926).
Third International Conference on Case Histories in Geotechnical Engineering
Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu

SOILS INVESTIGATION
Field Investigation
The field investigation consisted of a site
investigation conducted within days of the
collapse and a later investigation to obtain
.samples for soil testing. During soil sampling,
however, it was found that obtaining undisturbed
soil samples from the site was difficult. The
first difficulty was in obtaining undisturbed
samples in close proximity to the excavation. As
is shown in Figure 2, the building was located
within 6 ft (1.8 m) of an adjacent property and is
bounded by streets and a parking lot in the rear
of the building. The soils in the region of the
collapse were all highly disturbed due to the
collapse and could not be adequately sampled.
Consequently, soil samples were taken from the
rear of the building in the parking area. Six
auger holes were placed through the surface of the
parking lot. The second difficulty encountered
was that at about 40 in. (1 m) of depth very soft
satura1;ed soils were encountered. While the soils
~ere easily penetrated with a shelby tube, the
saturated soil would not remain in the tube upon
extraction.
In fact, the auger hole itself did
not stay open on account of the highly saturated
soils.
Undisturbed samples were, therefore,
obtained only for the top 40 in. of the soil
column. Auger holes were also placed in the area
of the excavation to confirm the presence of the
soft saturated soils.
These auger holes also
encountered very wet, saturated conditions with
standing water at about 40 in. (1m). However,
since the trench had considerable debris placed in
it after the collapse, this may have increased the
permeability of the backfilled trench area
allowing water to collect in this area.
Classification of the soils was based on the
following tests: Atterberg limits, grain size
analysis, natural water content, and unit weight
tests. strength measurements were obtained from
uniaxial compression tests. The foundation soils
were found to be a clayey silt CL-ML in the top 40
in. (1 m) tending to a low plasticity clay CL in
a lower zone of 40 to 80 in. (1 to 2 m).
The
liquid limits of the soils, ranged from 16 to 40%
while the plastic indexes were estimated to range
from 1. 5 to 18%. Void ratios for the undisturbed
samples varied between 0. 55 and o. 70. The natural
water content ranged from 15% near the top of the
soil column to 30% near the bottom of the column.
The unconfined compressive strength of the soils
in the upper 40 in. averaged 1060 psf (50 kPa).
At a depth of about 8 0 in. ( 2 m) , a very stiff
clay, difficult to auger through, was encountered
in all of the six auger holes completed. Figure
2 illustrates a cross-section of the building, its
foundation soils and the estimated location of the
exca.vation r~lative to the building.
Existing Soils Information
Since the foundations are relatively shallow and
the strip footings located at a depth of 18 in.
(457 mm), the USDA Soil Conservation Service soil
survey of Tuscaloosa county (1981) was reviewed
for additional information concerning the area's
soils. This survey had been issued in August of
1981, with the major field work being conducted
from 1971 to 1979.
Figure 3 below shows the
general soil map for the area of the building
collapse.
As can be seen from Figure 3, the
building is located at the contact of a soil type

1178

14ft

Remaining
Portion of
Building

Property

Line

Figure 2

Cross-section of co1lapse building site.

identified as a number 2 and described in the soil

survey as a Adaton silt loam. According to the
soil survey, the Adaton series consists of deep,
poorly drained, slowly permeable soils that formed
in silty fluvial sediments. In addition, the soil
survey provides information for planning land uses
related to urban development such as building site
ievelopment and engineering index properties. For
building development, the Adaton soils were
::lassified as being "severe" for building shallow
excavations and dwe1lings without basements. This
NaS primarily due to the soil's wetness and low
strength. Figure 4 lists the engineering index
properties of the Adaton soil series. It can be
seen from Figure 4 that the basic engineering
property tested
confirmed
the
field
test
previously reported.

footing unless the structure is first properly

underpinned or protected against settlement"
(Southern Standard Building Code, 1961). Based on
the soil strength measurements obtained, coupled
with the information provided in the USDA soil
survey of the area, it is highly unlikely that the
building's foundation soils would be able to
support a 5 to 6 ft excavation within 1 ft of the
foundation. To address this issue, the geological
setting of the foundation soils, a bearing
capacity analysis prior to the excavation, the
probable collapse mechanism, and compliance with
existing building codes are discussed.
Geologic Considerations
A significant feature of the soils in the
Tuscaloosa area is the alluvial orJ.gJ.n of the
unconsolidated terrace sediments deposited by the
Black warrior River. A well known characteristic
of alluvial deposits is the extreme vertical and
horizontal heterogeneity of the soils. due to
stream
migration
as
well
as
flood water
deposition.
The USDA Soil Conservation survey,
presented in Figure 3, shows that the building was
constructed at the edge of a pod-like section of
Adaton silt. However, soil sampling revealed that
a majority of the building, particularly the
section that collapse, was constructed over
approximately 7 ft (2 .1 m) of Adaton silt. Other
locations in the Tuscaloosa area in which the
Adaton silt is found are primarily in stream

WALYSIS
~ site investigation immediate1y following the
::ollapse found that the trench for the sewer line
1ad been placed within 1 ft of the foundation.
rhis was done due to the close proximity of the
ouilding to an adjacent property line, where only
. 6 ft ( 1. 8 m) strip of land was available for
'lacement of the sewer line.
Based on this
inding, the excavation was in violation of
.pplicable building codes, which require "that
:xcavations for any purpose near structures should
.ot extend within 1 ft (0.3 m) of the angle of
epose or natural slope of the soil under any

Third International Conference on Case Histories in Geotechnical Engineering

Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu

1179

Figure 3

Soil name and

map symbol

2---------------Adaton

Depth

I
I

Soil conservation Service soil survey.

Percentage paaslng
I
--...,,r:s.::ci:..:v:.::e__:nr,u"'m:::.be,.,r-=-;.;:~---1 L ~i~

Classification

USDA texture

Unified

I
....!1
''
''
0-7 !Silt
loam--------IML,
CL,
J

I CL-ML

7-961Silt loam, silty ICL, CH

! clay loam, siltyt
I clay.
I

Figure 4

AASHTO

'

'tA-4

100

IA-6, A-7

100

Lj 0

2 00

index

198-100190-100184-1001

<30

'I

I
I

I
I

'I

NP-10

I
I

198-100195-100184-1001 30-52
I

......_

1
I

10

'

Plasticity

11-30

Engineering index properties of the

Adaton
soil
series
(USDA,
Soil
Conservation Service).

channels.
This can be seen in Figure 3 where
Adaton silt is located to the southeast of the
building collapse area.
The pod-like nature of this soil suggests that
it was a low area where sediments collected over
time. This low area may also have been part of an
extended drainage system for receding flood waters
but over time filled with sediment. As is typical
of areas that accumulate sediment, the sediment
tends to be silt, which is a highly erodible soil,
along with some clays.
The results of the
Atterberg limits test indicate that the top
portion of the foundation soils are a low
plasticity silt that graded into a low plasticity
clay.
This would suggest that the foundation
soils formed over a long period of time from
surface erosion which would tend to collect more
silty than clayey soils, as opposed to resulting
directly from the receding flood waters. Since
this area is topographically a low area, it would
also tend to collect water. This was confirmE!d by
field observations in which the auger holes
encountered highly saturated soils in the lower
portion of the auger holes.
It appears that a
stiff clay, located at a depth of about 7 ft
(2m), trapped water above it.

Third International Conference on Case Histories in Geotechnical Engineering

Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu

Bearing Capacity Analysis

Although only limited strength data was obtained
from the site, an attempt was made to estimate the
bearing capacity of the foundation prior to the
excavation and subsequent building collapse. To
perform a bearing capacity analysis, an estimate
of the soil strength was made.
Based on the
unconfined compression test of the upper 40 in.
(0.1 m), the soil had an average unconfined
compressive strength of 1060 psf (50 kPa), giving
an undrained shear strength of 530 psf (25 kPa)
assuming ~=o. However, the lower portion of the
soil was not tested because undisturbed samples
were not obtained. Based on field observation and
.moisture content measurements, which ranged from
25 to 30% for the soils in the 40 to 80 in. zone,
it was assumed that the soil in this zone was
relatively close to its liquid limit. According
to Cassagrande (1932), soils near their liquid
limit have an undrained shear strength of
approximately 50 psf (2.4 kPa). This provides a
conservative estimate of the undrained shear
strength of the soil in the lower saturated soil.
Using these estimates of shear strength for the
1180

Collapse Mechanism

foundation soils, a bearing capacity analysis was

conducted assuming a strip footing supported by a
two-layer soil. A recently proposed semiempirical
equation by Azam and Wang (1991) for determining
the ultimate bearing capacity of a two-layer c-
soil was used in the analysis.
The proposed
bearing capacity equation is given as follows

Based on the above analysis and information

gathered at the site, the most probable collapse
mechanism was that once the trench was excavated,
the bottom soil, which was highly saturated and
possibly near its liquid limit, flowed into the
trench, thus undermining the upper bearing soils
and resulting in the collapse of the building.
This may also account for the time lag that
occurred between the completion of the trench and
the collapse, which was about 1 hour.
Azam and Wang also studied the plastic flow
behavior
of foundation soils using a finite
element technique to gain insight into the
progressive yielding of these soils.
Their
analysis of a stiff clay underlain by a weak clay
revealed that the yield zone extends deep into the
weaker bottom layer and that the yield pattern is
typical of a punching shear of the top layer
followed by a general shear failure of the bottom
layer.
Although
the
collapsed
building's
foundation soils were considerably weaker than
those studied by Azam and Wang, it is possible
that progressive yielding of the bottom soil also
accounted not only for the collapse but also the
time delay of the collapse.
Building Codes

where
q 0 = ultimate bearing capacity of a strip footing
over a two-layer soil;
q, = ultimate bearing capacity of the footing
supported by an infinitely thick top-layer
soil, computed by the traditional bearingcapacity equations using factors recommended
by Vesic (1975);
qb
ultimate bearing capacity of the footing
supported by an infinitely thick bottom-layer
soil, computed by the same method as q,;
m
layer factor, which is o .17 -o. 2 3 for twolayers of clay (use of the lower value is
recommended if one clay layer is highly
compressible);
H1 = distance between the base of the strip
footing and the top of the bottom layer;
B = width of the strip footing.

According to the standard building code section on

excavations, the excavation had been placed too
close to the foundation since it was not at least
1ft (0.3 m) from the angle of repose of the soil
or from the natural slope of the soil. Since the
"angle of repose" of a soil generally refers to a
granular material in loosely packed state, the
excavation angle would have to have been based on
the natural angle of the soil. However, this is
a difficult parameter to determine since the soils
were a soft saturated silty clay.
If it is
assumed that the soils were at or close to their
liquid limit, then the natural slope would be very
low and would result in large distances required
from the foundation for the excavation. Thus, the
only possible solution for this excavation would
have been to properly underpin or to prevent the
bottom soils from flowing into the trench, both of
which would have been difficult to complete given
the limited space available as well as the cost
involved.
An additional consideration in the collapse is
that although the contractor was in violation of
existing codes for excavation, their prior
experience with
other' excavations
in
the
Tuscaloosa area indicated that there should be no
problem with the excavation, since they had
excavated shallow trenches next to or near the
foundation without resulting in a collapse of the
trench. Therefore, no consideration was given to
first testing the soils to determine if they were
stable to excavate.
In addition, the backhoe
operator observed the lower soils in the trench
slowly flowing into the trench but yet did not
:consider the situation as dangerous.
While
training of the operator was lacking in this case,
information did exist from the USDA soil survey
for this area that was readily obtainable and that
would have indicated the possible dangers in
excavating these soils. In addition, knowledge of
the geologic origin of the soils in this area
would also help alert operators to the occurrence
of potentially difficult soils, since in this
case, the alluvial nature of the soils should be
relatively apparent given their location to the
Blac~ warrior River.

Azam and Wang analyzed four different layer

combinations with one of the combinations a stiff
clay underlain by a soft clay.
While the
overlying soil in this analysis cannot be
classified as stiff, as compared to the properties
of the soils tested by Azam and Wang, the analysis
was used in hopes of providing a lower bound
estimate of the bearing capacity of the foundation
prior to the excavation. In addition, instead of
computing the bearing capacity using factors
recommended by Vesic (1975), a lower bound
estimate of % and qb was made assuming a simple
uniaxial stress field below the footing which
gives %o = 2c (Chen and McCarron, 1991).
The
resulting bearing capacities are as follows

% = 1060 psf (51 kPa)

~

100 psf (5 kPa)

According to Azam and Wang's proposed equation,

the ultimate bearing capacity of the strip footing
over the two-layer soil is

% = 495 psf (23 kPa)

Since the estimated bearing load on the footing
was 325 psf (16 kPa), the foundation had a safety
factor of approximately 1.5 against bearing
capacity failure.
According to Vesic (1975),
however, situations in which safety factors are
less than 2.0 should be avoided.
Consequently,
l:he foundation was below a minimum recommended
safety factor prior to excavation of the trench.
rn addition, Vesic also recommends that "removal
:~f existing overburden by scour or excavation
sho~ld be given adequate consideration".
It is
ml~kely that this was contemplated at the time of
:onstruction of the building or that an analysis
ll'as performed.

Third International Conference on Case Histories in Geotechnical Engineering

Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu

1181

CONCLUSION
Conclusions concerning the building collapse can
be summarized as follows:
1.
2.
3
4.

5.

6.

7.

8.

The building had an estimated bearing

capacity SF ~ 1.5 prior to collpase.
Bearing soils were found to be a low
plasticity silty clay overlying a saturated
low plasticity clay.
A trench was placed within 1 ft ( 0. 3 m) of
the foundation and in violation of applicable
building codes for excavations.
Insufficient space was available between the
foundation and an adjacent property to place
the trench and meet existing codes for
excavations.
Collapse of the building is believed to have
resulted from the lower saturated clay
flowing into the trench, thus undercutting
the overlying soil and causing a loss of
bearing capacity.
No investigation of soil conditions was made
prior to the excavation nor were indications
of imminent failure of the trench taken into
consideration.
Information from the USDA Soil Conservation
Service soil survey showed that the building
probably was located on difficult soils which
may cause problems with excavations,
especially near foundations.
A significance of this collapse is that other
structures in the Tuscaloosa area are also
constructed on similar soils.
Since
information exists that can identify these
types of soils, it can be used to help
identify
foundation
soils
that
are
potentially susceptible to collapse from
adjacent shallow excavations.

Third International Conference on Case Histories in Geotechnical Engineering

Missouri University of Science and Technology
http://ICCHGE1984-2013.mst.edu

REFERENCES
Adams, G.I., Butts, c., Stephenson, L.W. and.W.
cooke (1926) "Geology of Alabama," Geolog~cal
Survey of Alabama, Special Report 14, p. 25.
Azam, G. and M.C. Wang (1991), "Bearing Capacity
of strip Footing Supported by Two-Layer c-
soil, Transportation
Rese~rch Rec~rd 1331,
TRB, National Research counc~l, Wash~ngton,
D.C., pp. 56-66.
cassagrande, A. (1932), "Rc;search on the Atterberg
Limits of Soils", Publ~c Roads, Vol. 13,
No. 8, pp. 121-136.
.
Chen, w.F. and w.o. McCarron (1991), "Bearu!-g
capacity of Shallow Foundations," Found-7t~on
Engineering Handbook, 2cd., ~. Y. Fang, ed~ tor,
Chapter 4, Van Nostrand Re~nhold, pp. 144-165.
Southern Standard Building Code, 1961, Southern
Building code Congress, Birmingham, AL, Section
1301, pp. 1-3.
USDA (1981), soil survey ~f Tusca~oosa County,
Alabama, soil Conservat~on Serv~ce and Forest
Service, pp. 118.
vesic A.S. (1963), "Bearing Capacity of Deep
Fou~dations
supported in Sand, " Highway
Research Record 39, HRB, National Research
council, Washington, D.C. pp. 112-143.
vesic, A.S. (1975), "Bearing Capacity of Shallow
Foundations," Foundation Engineering Handbook,
1st edn., H.F. Winterkorn and H.Y., Fang (eds.),
Chapter 3, Van Nostrand Reinhold Company, Inc.,
New York, N.Y.
wielchowsky, c.c. (1975), "Criteria fC?r
Distinguishing Pleistocene(?) Alluv~al Terrace
Deposits from the Coker Formation in ~he
cottondale, Alabama, Area, M.S. Thes~s
submitted to the University of Alabama, p. 238.

1182